Organisers: Eulàlia Bonet (UAB) and Maria-Rosa Lloret (U. of Barcelona)
Commentator:
Ricardo Bermúdez-Otero (U.
of Manchester)
PROGRAMME
Over the years,
opacity has been a major theme in phonological research, but, despite
the attention received, the issue is far from being settled. Opacity
effects refer either to surface forms that have undergone a process
although they do not match the criteria for its application at the
surface level (‘overapplication’), or to surface forms that fail to
undergo a process although their shapes match the criteria for it to
apply (‘underapplication’).
Opacity is predicted
by any theory allowing intermediate levels of analysis, that is, serial
theories: overapplication instances are cases of application of a
process at a non-surface level with the context for its application
being wiped out by a later process. Underapplication instances, instead,
are cases of non-application at the surface level that are controlled by
a non-surface level. Opacity, thus, constitutes an area where serial
approaches to phonology excel. Contrariwise, opacity effects are the
main challenge for parallel approaches to phonology and its resolution
often entails the introduction of otherwise unneeded devices. The
interest on opacity has nowadays revived because it questions the basic
architecture of Optimality Theory, which is generally assumed to be
global and parallel.
In serial modular
theories (such as Lexical Phonology) the opacity issue raises
interesting questions concerning the finding of independent motivation
for strata and the way of restricting the number of strata universally.
In parallel approaches, the most important issue raised by opacity —if
it is solved— is the theoretical cost of the devices introduced to
account for it. Within Optimality Theory, it further remains to be
investigated if a serial approach to Optimality Theory (such as OT-LP)
is a more restricted way of capturing opacity than its alternatives
(e.g., Sympathy Theory, Output-to-Output Correspondence, Local
Conjunction, Comparative Markedness, Candidate Chains). The implications
of both approaches to learnability and language typology also remain to
be seen. A further issue to be discussed is whether cases of opacity
really exist within one and the same level (or stratum), and whether
they can be dealt with under the auspices of either of the two competing
architectures or they call for a rule-based approach to phonology.
Finally, it has been shown that some alleged cases of opacity were not
really such when the data were reexamined. Since theories vary in the
range of opacity effects they predict, it is important to have opacity
phenomena well supported empirically.
Write to
cg.opacity@uab.es if you have any queries about the workshop.
